Hi Marcus,

>Everything at that level is determined entirely by the AD9361 RFFE
chip--although 
>there may be some settings that make things better.

It seems to make a difference if set_tx_bandwidth is called before or after
get_tx_stream.

In the example programs, they all seem to do it before, however, it appears
to work better afterwards. Please see the attached two programmes, which are
basically just your tx_samples_c program, except I've added in calls to
set_tx_bandwidth and set samples to 0.

I run with the args: -f 868300000 -g 89 -r 6000000

For lo_leakage_pre.c where bandwidth is set before get_tx_stream, carrier is
at -1dBm. For lo_leakage_post.c, where bandwidth is set to 56M before
get_tx_stream, then 6M afterwards, carrier is at -30dBm.

>Does the LO leakage have a strong frequency-dependent component? 

It doesn't appear to vary with frequency.

>I'd suggest getting yourself a copy of the full doc set for the AD9361
>-- ADI will provide them (not just the datasheet by other stuff) if
>   you "register" as a dev.

I'm reading this now, but I don't really know enough about UHD / FPGA
internals at the moment to know the order in which the various calibrations
are performed. I'd guess from the above, that the bandwidth may need to be
at a particular value for one of the calibration steps.

Thanks,
Jon



_______________________________________________
USRP-users mailing list
[email protected]
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com

Reply via email to